pharmaceutics

Article In Vitro Evaluation of Antimicrobial Activity of Formulations for Topical Application in Periodontal Therapy

Jan-Luca Schmid 1, Martin Kirchberg 2, Sandra Sarembe 3, Andreas Kiesow 3, Anton Sculean 4 , Karsten Mäder 2 , Mirko Buchholz 5 and Sigrun Eick 1,*

1 Laboratory of Oral Microbiology, Department of Periodontology, School of Dental Medicine, University of Bern, CH-3010 Bern, Switzerland; [email protected] 2 Institute of Pharmacy, Martin-Luther University Halle, D-06120 Halle (Saale), Germany; [email protected] (M.K.); [email protected] (K.M.) 3 Characterization of Medical and Cosmetic Care Products, Fraunhofer Institute for Microstructures and Materials IMWS, D-06120 Halle/Saale, Germany; [email protected] (S.S.); [email protected] (A.K.) 4 Department of Periodontology, School of Dental Medicine, University of Bern, CH-3010 Bern, Switzerland; [email protected] 5 Drug Design and Target Validation, Fraunhofer Institute for Cell Therapy and Immunology IZI-MWT and PerioTrap Pharmaceuticals GmbH, D-06120 Halle/Saale, Germany; [email protected] * Correspondence: [email protected]

 Received: 20 March 2020; Accepted: 11 April 2020; Published: 13 April 2020 

Abstract: Periodontal therapy using antimicrobials that are topically applied requires slow or controlled release devices. The in vitro antimicrobial activity of biodegradable polymer formulations that contain a new minocycline complex (P-MLC) was evaluated. The new P-MLC formulations that contained 11.5% minocycline were compared with pure minocycline or an existing commercial formulation, which included determination of minimal inhibitory concentration (MIC) values against two oral bacteria and activity on six-species periodontal biofilm. Moreover, the flow of gingival crevicular fluid (GCF) was modeled up to 42 days and the obtained eluates were tested both for MIC values and inhibiting biofilm formation. In general, MICs of the P-MLC formulations were slightly increased as compared with pure minocycline. Biofilm formation was clearly inhibited by all tested formulations containing minocycline with no clear difference between them. In 3.5 day old biofilms, all formulations with 250 µg/mL minocycline decreased bacterial counts by 3 log10 and metabolic activity with no difference to pure antimicrobials. Eluates of experimental formulations showed superiority in antimicrobial activity. Eluates of one experimental formulation (P503-MLC) still inhibited biofilm formation at 28 days, with a reduction by 1.87 log10 colony forming units (CFU) vs. the untreated control. The new experimental formulations can easily be instilled in periodontal pockets and represent alternatives in local antimicrobials, and thus warrant further testing.

Keywords: controlled release; periodontitis; local ; in vitro model; gingival flow

1. Introduction Periodontal disease is associated with an imbalance of microorganisms in subgingival biofilm. Bacteria highly present in these biofilms are Porphyromonas spp. and Tannerella forsythia [1]. Porphyromonas gingivalis has been designated as a keystone pathogen by modifying the biofilm towards a pathogenic one via modulation of the host response [2]. The response of the periodontium to inflammatory changes resulted in the loss of attachment and periodontal integrity [3].

Pharmaceutics 2020, 12, 352; doi:10.3390/pharmaceutics12040352 www.mdpi.com/journal/pharmaceutics Pharmaceutics 2020, 12, 352 2 of 14

Removal of the subgingival biofilm, either by manual or ultrasonic instrumentation, is an essential outcome of periodontal therapy and decreases the bacteria count associated with periodontal disease. When applied as a sole method, failures arise with increasing probing depths ( 5 mm) [4]. Systemic ≥ antibiotics improve the clinical outcomes, although their use should be restricted to severe cases [5], as antimicrobial use affects the gut microbiota [6] and is clearly associated with global development of antimicrobial resistance [7]. In particular, in supportive periodontal therapy, local antimicrobials are applied to treat residuals pockets [8]. Topical application can lead to a high concentration of the antimicrobial which is active in the diseased area and selectively targets microorganisms in the diseased area [9,10]. Systemic side effects can be reduced and application of controlled compliance by the periodontist to the patient is guaranteed [9,10]. However, an important problem is the high flow rates that are 20 µL/h of the gingival crevicular ≥ fluid (GCF) [11]. In order to avoid rapid depletion of the therapeutics and to keep them in place, sustained or controlled release devices are needed [9]. Desirable properties of the devices include easy application, biodegradability, and a continuous release of an antimicrobial-active substance over a defined time [9,10]. In so doing, topical antimicrobials combined with mechanical therapy can improve the clinical outcome [8]. This is underlined by a systemic review including more than 50 studies, in which the effect was most pronounced in deep and residual pockets [12]. Frequently, and analogs are used as local antibiotics [12]. Tetracylines (e.g., , minocycline, and ) block bacterial growth by inhibiting protein synthesis [13]. Moreover, independent of the antimicrobial activity, they are frequently used to inhibit matrix metalloproteinases activities [14]. Several formulations are on the market, such as gels containing doxycycline [15,16] or PLGA (poly-(lactide-co-glycolide)) based minocycline microspheres [17,18]. For the application of minocycline microspheres, the cartridge is connected with a special handle mechanism and used to insert the unit-dose cartridge into the base of the periodontal pocket, then, the thumb ring in the handle mechanism is pressed to expel the powder [19]. The minocycline microspheres have been proven in several clinical studies where the adjunctive used for mechanical debridement was compared with mechanical debridement only. A large clinical trial including more than 700 patients found minocycline microspheres highly efficient in probing depth reduction up to nine months [18]. Another study that included 127 subjects showed favorable results on the reduction of P. gingivalis, T. forsythia, and Treponema denticola [20]. Moreover, the adjunctive use of minocycline microspheres was able to improve the clinical outcome in the treatment of peri-implantitis [21]. In the present study, we investigated new experimental biodegradable formulations in the form of the extrudates, which could easily be placed into a periodontal pocket using a pair of tweezers. Recently, we described the manufacturing of new lipophilic complexes of tetracyclines, and their incorporation into these extrudates (Figure1)[22]. The new formulations show higher stability in an aqueous environment as compared with the pure drug molecules [22]. This is an important aspect, as drug molecules in biodegradable PLGA polymers could be exposed to a very acidic aqueous microenvironment (pH 2) for longer times prior to their release [23]. This exposure could lead to drug inactivation by or the formation of covalent amide bonds [24]. The main focus of the first publication was on the complex formation, the analytical characterization, and the drug release kinetics from polymer-lipid complex extrudates. Although the principal bioactivity of the lipophilic complexes has been shown in preliminary experiments [22], much more detailed studies are necessary to explore the potential of the new formations with regard to its biological performance. Therefore, this study aimed to investigate the time dependence of the antimicrobial activity in relevant in vitro models. In particular, the dilution by the GCF flow was mirrored. The newly developed polymer-lipid complex extrudates were compared with a commercially available product and minocycline as the drug substance. Pharmaceutics 2020, 12, 352 3 of 14 Pharmaceutics 2020, 12, x 3 of 14

Figure 1.Figure The 1. investigatedThe investigated minocycline minocycline lipid lipid complex complex (P-MLC)(P-MLC systems) systems are manufactured are manufactured in two in two sequentialsequential main steps. main steps.Step Step1 includes 1 includes the the formation formation of thethe lipophilic lipophilic minocycline minocycline lipid complex lipid complex MLC MLC by solvent and heat associated complex formation between minocycline and magnesium stearate. After by solvent and heat associated complex formation between minocycline and magnesium stearate. evaporation of the intermediate solvent ethanol, the MLC complex is micronized, mixed with the After evaporationbiodegradable of polymerthe intermediate poly-(lactide- solventco-glycolide) ethanol, (PLGA), the and MLC extruded. complex As a result,is micronized, a polymer-lipid mixed with the biodegradablecomplex extrudate polymer P-MLC poly is-(lactide obtained- (diameterco-glycolide) 600 µ m).(PLGA Two), di andfferent extruded. PLGA polymers As a result, were used a polymer- lipid complexP502-MLC extrudate and P503 -MLC.P-MLC is obtained (diameter 600 µm). Two different PLGA polymers were used2. P Materials502-MLC and and P Methods503-MLC.

2. Materials2.1. Antimicrobials and Methods The manufacturing and characterization of the PLGA and minocycline lipid complex (P-MLC) 2.1. Antimicrobialsextrudates have been described in a previous publication [22]. In short, for both used formulations (P502-MLC and P503-MLC), minocycline (Ontario Chemicals Inc., Guelph, ON, Canada) was chelated Thewith manufacturing calcium and preferably and characterization magnesium stearate of the (both PLGA Magnesia and Germany,minocycline Müllheim, lipid Germany) complex in (P-MLC) extrudatesthe ha molarve been ratio 1:2.described Subsequently, in a theprevious complex publication was mixed with [22]. the In desired short, PLGA for both polymer used (Evonik, formulations (P502-MLCDarmstadt, and P503 Germany)-MLC), minocycline and cryomilled. (Ontario This composition Chemicals was utilized Inc., Guelph, for the hot ON, melt Canada extrusion) withwas a chelated with calcium600 µm and device. preferably Each of the magnesium extrudates contained stearate 11.5%(both (m Magnesia/m) of minocycline Germany, [22]. Müllheim, Germany) in Positive controls were minocycline as a substance (Ontario Chemicals Inc., Ontario, ON, Canada) the molarand ratio minocycline 1:2. Subsequently, microspheres (Arestin the complex®; OraPharma, was mixed Bridgewater, with NJ, the USA). desired As a negativePLGA polymer control, the (Evonik, Darmstadt,drug Germany) free PLGA/lipid and extrudate cryomilled. (P503 and This of magnesiumcomposition stearate; was vehicle)utilized was for included. the hot All melt minocycline extrusion with a 600 µmformulations device. Each were of also the kept extrudates thoroughly contained in the dark 11.5 throughout% (m/m the) of experiments. minocycline [22]. PositiveTo controls achieve a homogeneous were minocycline distribution as of a the substance formulations (Ontario in the assays, Chemic the extrudatesals Inc., andOntario, the ON, microspheres were kept at 20 C overnight in a freezer, and thereafter micronized with a pestle in a Canada) and minocycline microspheres− ◦ (Arestin® ; OraPharma, Bridgewater, NJ, USA). As a negative porcelain mortar. control, the drug free PLGA/lipid extrudate (P503 and of magnesium stearate; vehicle) was included. All minocycline2.2. Microorganisms formulations were also kept thoroughly in the dark throughout the experiments. To achieveIn the a homogeneous assays, the following distribution six bacterial of the strains formulations were used: inPorphyromonas the assays, the gingivalis extrudatesATCC and the microspheres33277, wereTannerella kept forsythia at −20ATCC °C overnight 43037, Fusobacterium in a freezer nucleatum, and thereafterATCC 25586, micronizedStreptococcus with gordonii a pestle in a porcelainATCC mortar. 10558, Actinomyces naeslundii ATCC 12104, and Parvimonas micra ATCC 33270. All strains were precultured on tryptic soy agar plates (Oxoid, Basingstoke, GB) with 5% sheep blood for 24 h at 37 ◦C with the respective conditions (S. gordonii, A. naeslundii with 10% CO , others anaerobically). 2.2. Microorganisms 2 The determinations of minimal inhibitory concentration (MIC) values were made only against In theP. gingivalis assays, ATCCthe following 33277 and sixS. gordonii bacterialATCC strains 10558. were used: Porphyromonas gingivalis ATCC 33277, Tannerella forsythia ATCC 43037, Fusobacterium nucleatum ATCC 25586, Streptococcus gordonii ATCC 10558, Actinomyces naeslundii ATCC 12104, and Parvimonas micra ATCC 33270. All strains were precultured on tryptic soy agar plates (Oxoid, Basingstoke, GB) with 5% sheep blood for 24 h at 37 °C with the respective conditions (S. gordonii, A. naeslundii with 10% CO2, others anaerobically). The determinations of minimal inhibitory concentration (MIC) values were made only against P. gingivalis ATCC 33277 and S. gordonii ATCC 10558. All six strains were included in the biofilm assays. For that, bacteria were suspended in phosphate buffered saline (PBS) according McFarland 0.5 (about 1.5 × 108 microorganisms) and mixed in a ratio 1:2:4 (S. gordonii)/(A. naeslundii)/(each other strain).

2.3. Determination of Minimal Inhibitory Concentration (MIC)

Pharmaceutics 2020, 12, 352 4 of 14

All six strains were included in the biofilm assays. For that, bacteria were suspended in phosphate buffered saline (PBS) according McFarland 0.5 (about 1.5 108 microorganisms) and mixed in a ratio × 1:2:4 (S. gordonii)/(A. naeslundii)/(each other strain).

2.3. Determination of Minimal Inhibitory Concentration (MIC) A two-fold dilution series of the formulations dissolved in Aqua Dest was prepared. The highest concentration of the formulations was always equivalent to 64 µg/mL minocycline. The vehicle was adapted to the amount of this concentration of the test formulation. Thereafter, each 100 µL of test substance solution followed by 100 µL of bacterial suspension (about 106 bacteria/mL) in doubled concentrated Wilkins–Chalgren broth (Oxoid) with 10% of lysed horse blood and 10 µg/mL β-NAD (Merck KGaA, Darmstadt, Germany) was pipetted per well of a 96-well-plate. The final tested concentrations of the formulations ranged from 32 µg/mL to 0.125 µg/mL minocycline. After an incubation time of 18 h (anaerobes 42 h) in respective atmosphere, the MIC was assessed as the lowest concentration without visible growth (turbidity).

2.4. Activity on Biofilm Formation First, the test formulations were dissolved in Aqua Dest equivalent to 125 µg/mL. Then, the wells of a 96-well-plate were coated with 10 µL/well for 1 h, before 10 µL/well PBS with 1.5% bovine serum albumin (PBS/SA) was added for 10 min followed by the mixed bacterial suspension in Wilkins–Chalgren broth with 5% of lysed horse blood and 5 µg/mL β-NAD. Here, the total counts of colony forming units (CFU) of the biofilms were counted after 6 h and 24 h incubation in an anaerobic atmosphere at 37 ◦C. After 24 h of incubation, the quantity and metabolic activity of the biofilms were also measured. The quantification was made by staining with crystal violet according to recently published protocols [25,26]. In short, after washing, the biofilms were fixed at 60 ◦C for 1 h. Thereafter, 50 µL of 0.06% crystal violet solution was given per well. After 10 min of incubation at room temperature, the stained biofilms were dissolved by 200 µL of 30% , and after transferring to a new microtiter plate, the OD600nm was measured. Biofilm metabolic activity was assessed using resazurin as a indicator [25,27]. After short washing of the biofilms, 100 µL of nutrient broth containing 0.06 mM resazurin (Merck KGaA) per well was added. After 1 h of incubation at 37 ◦C, the plate was measured at 570 nm against 600 nm.

2.5. Activity on an Already Formed Biofilm Biofilms using the multispecies mixture consisting of 6 species were formed for four days. Nutrient broth was exchanged and selected bacteria (P. gingivalis, T. forsythia) were added again after 2.5 days. At four days of biofilm formation, after removing the nutrient media and short washing with PBS, 100 µL of test substances solutions in a concentration equivalent to 1000, 500, and 250 µg/mL were added per well for 10 min, before 100 µL of nutrient broth were pipetted per well (final concentrations of minocycline 50, 250, and 125 µg/mL). After an overnight incubation (18 h), the CFU, biofilm quantity, and biofilm metabolic activity were determined, as described before. The Biofilm experiments were made in two independent experiments in each independent quadruplicate. ANOVA with post hoc Bonferroni was used for statistical analysis. The level of significance was set to p = 0.05. Software SPSS 25.0 (IBM SPSS Statistics, Chicago, IL, USA) was used.

2.6. Simulation of the Release Kinetic in a Periodontal Pocket The GCF flow was simulated in a model. The basis of the experimental setup was a resting volume of 1.5 µL with a flow of 44 µL/h in periodontitis, which decreased to 15 µL/h 12 weeks after therapy (Goodson et al., 2003). Thus, formulations equivalent to 1 mg minocycline were put into tubes with 23.5 µL PBS/SA. After 30 min, tubes were centrifuged at 5000 g for 1 min and 22 µL were removed × and replaced by fresh 22 µL PBS/SA. Then, at diverse time points, the exchange of the eluates using different volumes was repeated (Table1). All obtained eluates were stored at 20 C until analysis. − ◦ Pharmaceutics 2020, 12, 352 5 of 14

Table 1. Scheme of pipetting to simulate the gingival flow.

PBS/SA Per Tube with Antimicrobial (1 mg) Time Removal (µL) Addition (µL) T 0 - 23.5 T 30 min 22 22 T 60 min 22 44 T 2–4 h each h 44 44 T 4 h 44 88 T 6 h 44 792 T 24 h 792 1056 T 2 d–T 3 day each day 1056 1056 T 4 day 1056 3168 T 7 days 3168 3500 T 10.5 day 3500 3500 T 14 days 3500 3304 T 17.5 day 3304 3304 T 21 days 3304 3108 T 24.5 day 3108 3108 T 28 days 3108 2912 T 31.5 day 2912 2912 T 35 day 2912 2716 T 39.5 day 2716 2716 T 42 days 2716 -

For MIC determination, eluates were used which were obtained at 1, 2, 4, 6, 24 h, and 2, 4, and 7 days, and then weekly up to 42 days. A two-fold dilution series of the eluates was prepared. Eluates were handled as the dissolved formulations before in the assays. The highest dilution of the eluate inhibiting growth of S. gordonii and P. gingivalis was recorded. Furthermore, to assess the potential activity on the biofilm formation, eluates obtained at 24 h, and 2, 7, 14, and 28 days were used. The wells of a 96-well plates were coated with 10 µL eluate/well for 1 h. The further procedure of biofilm formation was as mentioned before. However, here, the CFU were counted after only 6 h of incubation.

3. Results

3.1. Antimicrobial Activity of the Formulations Against Planktonic Bacteria The minimal inhibitory activity of the P-MLC formulations was slightly less, but in the range of the pure substance. The difference was one to two dilution steps related to the pure substance. The microspheres were equally active as the substance. The minocycline free vehicle did not exert any antibacterial activity against the tested bacteria (Table2).

Table 2. Minimal inhibitory concentrations of the minocycline formulations (equivalent to µg/mL minocycline).

Formulation S. gordonii ATCC 10558 P. gingivalis ATCC 33277 Minocycline 0.5 0.25 Microspheres 0.5 0.25 P502-MLC 1 0.25 P503-MLC 2 0.5 Vehicle No inhibition No inhibition

3.2. Activity of the Formulations on Biofilms All tested experimental formulations and positive controls clearly inhibited biofilm formation (Figure2). Regarding the CFU counts, the pure substance was the most active, with a reduction Pharmaceutics 2020, 12, x 6 of 14

All tested experimental formulations and positive controls clearly inhibited biofilm formation

(FigurePharmaceutics 2). Regarding2020, 12 ,the 352 CFU counts, the pure substance was the most active, with a reduction6 of 14 of 2.1 log10 after 6 h and 3.6 log10 after 24 h. The biofilm quantity was about one-sixth and the metabolic activity was about one-sixth, one-tenth related to the controls, here, no difference among the of 2.1 log10 after 6 h and 3.6 log10 after 24 h. The biofilm quantity was about one-sixth and the formulationsmetabolic with activity minocycline was aboutone-sixth, occurred one-tenth, but the related vehicle to the also controls, reduced here, nothe di ffCFUerence counts among by the 0.6 log10 and theformulations biofilm quantity with minocycline by about occurred, one-third. but the vehicle also reduced the CFU counts by 0.6 log10 and the biofilm quantity by about one-third.

Figure 2. Activity of the minocycline formulations (substance (minocycl.) and microspheres (microsph.) Figure polymer-lipid 2. Activity complex of the extrudate minocycline (PLM-C)) formulations and PLM-C without (substance minocycline (minocycl.) (vehicle) on formationand microspheres of (microsph.)a six-species polymer biofilm.-lipid Bacterial complex counts extru determineddate (PLM as colony-C)) and forming PLM units-C without (CFU) (A );minocycline biofilm quantity (vehicle) on formation(B); andof a metabolic six-species activity biofilm (C). **. (black)Bacterialp < 0.01 counts vs. control; determined ** (blue) asp < colony0.01 vs. Pforming502-MLC; units * (green) (CFU) (A); p < 0.05 / ** (green) p < 0.01 vs. P503-MLC. biofilm quantity (B); and metabolic activity (C). ** (black) p < 0.01 vs. control; ** (blue) p < 0.01 vs. P502-

MLC;3.3. * Activity (green) on p an< 0.05 Already / ** Formed(green) Biofilm p < 0.01 vs. P503-MLC. Analyzing the activity on an already formed biofilm, the reduction of the CFU counts was between 3.3. Activity2.4 and on 3.7 an log10 Already CFU F withormed a tendencyBiofilm of more activity with the higher concentrated minocycline formulations. Among the minocycline formulations, there was no difference and metabolic activity Analyzingwas also clearly the activity reduced byon all an minocycline already formed formulations. biofilm, No clear the ereductionffect on biofilm of the quantity CFU was counts was betweenvisible. 2.4 and The vehicle 3.7 log10 did not CFU show with any activity a tendency on the established of more biofilm activity (Figure with3). the higher concentrated minocycline formulations. Among the minocycline formulations, there was no difference and metabolic activity was also clearly reduced by all minocycline formulations. No clear effect on biofilm quantity was visible. The vehicle did not show any activity on the established biofilm (Figure 3).

Pharmaceutics 2020, 12, 352 7 of 14 Pharmaceutics 2020, 12, x 7 of 14

FigureFigure 3. 3.ActivityActivity ofof the the minocycline minocycline formulations formulations (substance (substance (minocycl.) (minocycl.) and microspheres and (microsph.)microspheres (microsph.)polymer-lipid polymer complex-lipid extrudate complex (PLM-C)) extrudate and (PLM PLM-C-C)) without and PLM minocycline-C without (vehicle) minocycline on a six-species (vehicle) on a sixbiofilm-species formed biofilm over formed 3.5 d. Bacterialover 3.5 countsd. Bacterial determined counts as determined colony forming as colony units (CFU)forming (A); units biofilm (CFU) (A)quantity; biofilm ( Bquantity); and metabolic (B); and activity metabolic (C). activi ** p < ty0.01 (C vs.). ** control. p < 0.01 vs. control. 3.4. Antimicrobial Activity of the Eluates on Planktonic Bacteria 3.4. Antimicrobial Activity of the Eluates on Planktonic Bacteria Initially, eluates of all minocycline formulations inhibited the growth of S. gordonii ATCC 10558 andInitially,P. gingivalis eluatesATCC of all 33277. minocycline As expected, formulations the activity inhibited decreased the continuously growth of over S. gordonii time. At ATCC 28 days, 10558 andthere P. gingivalis was no remaining ATCC 33277. activity As for expected, the pure drug the substanceactivity decreased and the microspheres, continuously whereas over thetime. eluates At 28 d, thereof the was P-MLC no remaining were still active activity after for 42 thedays. pure The drugeluates substance of the vehicle and did the not microspheres, have any activity whereas at any the eluatestime of (Figure the P4-).MLC were still active after 42 d. The eluates of the vehicle did not have any activity at any time (Figure 4).

Pharmaceutics 2020, 12, 352 8 of 14

Figure 4. Minimal inhibitory concentration (maximum dilution) of the eluates obtained over a period of 42 days from minocycline formulations (substance (minocycl.) and microspheres (microsph.) polymer-lipid complex extrudate (PLM-C)) and PLM-C without minocycline (vehicle) according to 1 µg of minocycline and simulating the flow of the gingival fluid against Streptococcus gordonii ATCC 10558 (A) and Porphyromonas gingivalis ATCC 33277 (B).

Considering the MIC value of the minocycline substance, the antimicrobial active concentration of minocycline was calculated. At 1 h, it was >16 mg/mL for the substance and the microspheres and 2 to 8 mg/mL for the P-MLC. At 24 h, the values decreased to 62.5 µg/mL for the substance, each 125 µg/mL for the P502-MLC formulation and the microspheres, and 250 µg/mL for the P503-MLC formulation. From 21 days, this concentration was less than 1 µg/mL for the substance and the microspheres, whereas the eluates of both P-MLC formulations had an activity of 4 µg/mL minocycline (Table S1).

3.5. Antibiofilm Activity of the Eluates All eluates obtained from minocycline preparations at 1 day inhibited biofilm formation by about 2 log10 CFU. Later, up to 14 days, eluates only obtained from minocycline P-MLC or microspheres formulations reduced the CFU in biofilms by more than 1 log10. At 28 days, only the eluate obtained from P503-MLC showed a reducing effect (1.9 log10) and in addition, the eluates obtained from the vehicle decreased the CFU in biofilm by 1.4 log10 and 1.1 log10 at day 14 and day 21, respectively (Figure5).

1

PharmaceuticsPharmaceutics 20202020,, 1122,, x 352 9 of9 14of 14

Figure 5. Activity of the eluates obtained over a period of 42 days from minocycline formulations Figure(substance 5. Activity (minocycl.) of the and eluates microspheres obtained (microsph.) over a period polymer-lipid of 42 d complex from minocycline extrudate (PLM-C)) formulations and (substance (minocycl.) and microspheres (microsph.) polymer-lipid complex extrudate (PLM-C)) and PLM-C without minocycline (vehicle) according to 1 µg of minocycline and simulating the flow of the PLM-C without minocycline (vehicle) according to 1 µg of minocycline and simulating the flow of the gingival fluid on biofilm formation (CFU after 6 h). ** (black) p < 0.01vs. control; ** (blue) p < 0.01 vs. gingival fluid on biofilm formation (CFU after 6 h). ** (black) p < 0.01vs. control; ** (blue) p < 0.01 vs. P502-MLC; ** (green) p < 0.01 vs. P503-MLC. P502-MLC; ** (green) p < 0.01 vs. P503-MLC. 4. Discussion and Conclusions 4. Discussion and Conclusions In this study, newly developed formulations with magnesium stearate and PLGA as the vehicle wereIn evaluated this study, regarding newly developed their antimicrobial formulations activity. with Except magnesium for analyzing stearate the and formulations PLGA as the directly, vehicle werethe gingival evaluated flow regarding was simulated, their antimicrobial and eluates of activity. the formulations Except for obtained analyzing for athe period formulations of 42 days directly, were theinvestigated. gingival flow A controlled was simulated long-lasting, and activity eluates of of the the minocycline formulations as the obtained loaded antibioticfor a period was of confirmed. 42 d were investigated.The formulations A controlled were long prepared-lasting as activity extrudates of with the minocycline a 600 µm device, as the of whichloaded the antibiotic P503-MLC was confirmed.extrudate had a larger diameter than the P502-MLC extrudate due to viscoelastic properties [22]. For the experiments,The formulations they were were available prepared as cylinders as extrudates and easy with to a cut 600 to µm the necessarydevice, of size.which However, the P503-MLC for extrudateclinical use, had other a larger sizes diameter and shapes than could thealso P502- beMLC created. extrudate due to viscoelastic properties [22]. For the experiments,Minocycline they was selectedwere available because as of cylinders the fact that and it easy is one to of cut the to best the documented necessary size. local However, antibiotics for clinicalin periodontal use, other therapy sizes and [12]. shapes In general, could minocycline also be created. is also discussed as an anti-inflammatory drug for theMinocycline treatment was of Alzheimer’s selected because disease of the [28 ]fact and that a cardiovascular it is one of the best therapeutic documented agent local related antibiotics to its inanti-inflammatory, periodontal therapy antiapoptotic, [12]. In general, and antioxidant minocycline properties is also [discussed29]. as an anti-inflammatory drug for theAs treatment a local drug of inAlzheimer’s periodontal disease therapy, [28] diff anderent a approachescardiovascular have therapeutic been taken, agent such asre loadedlated to in its antiPLGA-inflammatory, hydrogels [30 antiapoptotic,], in in situ reversed and antioxidant hexagonal properties mesophase [29] [31.], or in liquid crystals [32]. In this study,As the a local experimental drug in periodontal formulations therapy were compared, different with approaches minocycline have microspheres been taken, which such as have loaded been in commercially available for many years. PLGA hydrogels [30], in in situ reversed hexagonal mesophase [31], or in liquid crystals [32]. In this First, experimental formulations were directly compared with pure minocycline and minocycline study, the experimental formulations were compared with minocycline microspheres which have microspheres. A comparison of the different formulations showed that the pure substance was the been commercially available for many years. most active against the two tested species. The experimental formulations exerted slightly less First, experimental formulations were directly compared with pure minocycline and activity. The purpose of a controlled release device is that the antimicrobial activity is prolonged over minocycline microspheres. A comparison of the different formulations showed that the pure substance was the most active against the two tested species. The experimental formulations exerted slightly less activity. The purpose of a controlled release device is that the antimicrobial activity is

Pharmaceutics 2020, 12, 352 10 of 14 time. Although the experimental formulations were dispersed as much as possible, it is suggested that minocycline was still covered by the vehicle substances. The total release of minocycline from these formulations was found to be only half of the minocycline magnesium and calcium stearate complexes [22]. Biofilm formation findings confirm the MIC results regarding the viable bacterial counts within the biofilm. Applying the formulations on an already formed biofilm, there was no visible difference between the formulations. Viable bacterial counts decreased by about 3 log10, but a biofilm eradication did not occur, although using high concentrations of the antibiotic were used over time. Once more, this finding underlines that antibiotics alone are not able to eliminate bacteria within the biofilm [33], and therefore, in periodontal therapy, their use is limited as adjuncts to instrumentation [34]. To mimic release kinetics in in vitro assays, in most cases the controlled release device is placed in defined, constant volumes over time [35–37]. In order to simulate the clinical situation as closely as possible, a GCF flow during a periodontal therapy with decreasing volumes [11] was simulated. Thus, the initial volume was adjusted to those of the periodontal pocket and the exchange was simulated accordingly taking into account the reduced GCF flow during periodontal therapy. Regarding its protein composition, since the GCF closely resembles a diluted serum [38], a buffered saline solution supplemented with serum albumin was added. The obtained eluates were used for analysis of the antimicrobial activity, up to 42 days. However, limitations of the in vitro assay have to be addressed. A short centrifugation step was included mainly for the purpose of sedimenting condensed in the tubes. In particular, at the beginning of the experiments (when having only small volumes), an effect on the release of the minocycline into the environment cannot be completely excluded. In addition, the method did not consider a potential attachment and uptake of the antibiotics by epithelial cells, fibroblasts, and immune cells. Minocycline uptake by oral epithelial cells has been reported by [39]. Further accumulation in gingival fibroblasts has been shown [40]. Tetracycline analogs seem to be adsorbed mainly by plasma proteins and, when administered orally, their concentrations in GCF were found to be low, in about 50% they did not reach 1 µg/mL [41]. However, the concentration of tetracycline can reach more than 1000 µg/mL in the periodontal pocket when applied as a non-resorbable fiber over a few days [42]. For doxycycline, another tetracycline analogous, the concentrations in the periodontal pocket were about 1000 µg/mL after 2 h of application of the formulations containing either 8.5% or 14% of doxycycline, and then the concentrations decreased to 8 and 19 µg/mL after 12 days [43]. In the present study, the in vitro antimicrobial activity of the newly developed formulations lasted up to 42 days, while the activity of the minocycline microspheres was detectable only up to 21 days. Unfortunately, concerning the concentration or antimicrobial activity of minocycline, after applying minocycline microspheres in the periodontal pockets, exact data are not available. It is only mentioned once that the concentration was about 340 µg/mL in the GCF 14 days after application [18]. Differences between the eluates of the two experimental formulations were seen regarding their direct inhibitory activity on bacterial growth, as well as on retarding biofilm formation. The P503-MLC was more active than the P502-MLC. The higher activity could be related to the larger diameter [22] and the resulting higher surface which allows an initial faster release of the antibiotic. The biofilm experiments showed that the vehicle itself (both directly and the obtained eluate) was able to inhibit the biofilm formation. The vehicle contained PLGA and magnesium stearate. PLGA is often incorporated in local antibiotics formulations [44,45]. However, a direct antimicrobial activity of the vehicle was not found which is in accordance with studies using PLGA nanoparticles [45,46]. It has also been reported that unloaded PLGA-nanoparticles did not affect biofilm formation of Pseudomonas aeruginosa [46,47]. It is suggested that the inhibition of biofilm formation is linked to the incorporated magnesium stearate. In particular, magnesium is being discussed as an implant material, it exerts antibacterial activity against planktonic Staphylococcus epidermidis, Staphylococcus aureus, and Escherichia coli, furthermore, it prevents biofilm formation of these species [48]. In addition, magnesium oxide nanoparticles clearly inhibited the growth of Staphylococci, E. coli, Pseudomonas aeruginosa, and Candida ssp. in a study by [49] which also demonstrated a disruption of a nascent biofilm by magnesium oxide nanoparticles [49]. Pharmaceutics 2020, 12, 352 11 of 14

In conclusion, the tested controlled release device loaded with minocycline (especially P503-MLC) seems to be a suitable alternative to commonly used systems, thus warranting further testing in animal models, and thereafter in clinical trials.

5. Patents The hereby presented work was filed in a patent application, which was registered under PCT/EP2019/079566 at the EPA dated on 2019/10/29 as a community invention of Fraunhofer-Society, Martin-Luther-University Halle-Wittenberg and the University of Bern.

Supplementary Materials: The following are available online at http://www.mdpi.com/1999-4923/12/4/352/s1, Table S1: Antimicrobially active concentration of minocycline (µg/mL) in eluates obtained from minocycline P-MLC or microspheres formulations over a period of 42 days. Author Contributions: Conceptualization, A.K., M.B., and K.M.; methodology, K.M. and S.E.; formal analysis, J.-L.S., S.S., and M.B.; investigation, J.-L.S. and M.K.; resources, S.S., K.M., and A.S.; data curation, K.M., A.K., and S.E.; writing—original draft preparation, J.-L.S., M.K., S.S., and A.K.; writing—review and editing, A.S., K.M., M.B., S.E.; visualization, M.K., S.S., and S.E.; supervision, K.M. and S.E.; project administration, A.K.; funding acquisition, A.K., M.B., and K.M. All authors have read and agreed to the published version of the manuscript. Funding: This research was funded by a grant (EU-EFRE; IB: 1604/00078) of the German state Saxony-Anhalt. Acknowledgments: The authors are grateful to Prashanthnj Sivapatham and Anna Magdo´n(Dental Research Center, Laboratory of Oral Microbiology, University of Bern) for their excellent assistance in performing the in vitro assays. Conflicts of Interest: Mirko Buchholz is CSO and holds shares of PerioTrap Pharmaceuticals GmbH, a spin-off company of Fraunhofer IZI-MWT. The other authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

References

1. Lourenco, T.G.; Heller, D.; Silva-Boghossian, C.M.; Cotton, S.L.; Paster, B.J.; Colombo, A.P. Microbial signature profiles of periodontally healthy and diseased patients. J. Clin. Periodontol. 2014, 41, 1027–1036. [CrossRef] 2. Hajishengallis, G.; Lamont, R.J. Breaking bad: Manipulation of the host response by Porphyromonas gingivalis. Eur. J. Immunol. 2014, 44, 328–338. [CrossRef][PubMed] 3. Kurgan, S.; Kantarci, A. Molecular basis for immunohistochemical and inflammatory changes during progression of gingivitis to periodontitis. Periodontol 2000 2018, 76, 51–67. [CrossRef] 4. Cobb, C.M. Clinical significance of non-surgical periodontal therapy: An evidence-based perspective of scaling and root planing. J. Clin. Periodontol. 2002, 29 (Suppl. S2), 6–16. [CrossRef][PubMed] 5. Pretzl, B.; Salzer, S.; Ehmke, B.; Schlagenhauf, U.; Dannewitz, B.; Dommisch, H.; Eickholz, P.; Jockel-Schneider, Y. Administration of systemic antibiotics during non-surgical periodontal therapy-a consensus report. Clin. Oral Investig. 2019, 23, 3073–3085. [CrossRef][PubMed] 6. Palleja, A.; Mikkelsen, K.H.; Forslund, S.K.; Kashani, A.; Allin, K.H.; Nielsen, T.; Hansen, T.H.; Liang, S.; Feng, Q.; Zhang, C.; et al. Recovery of gut microbiota of healthy adults following antibiotic exposure. Nat. Microbiol. 2018, 3, 1255–1265. [CrossRef] 7. Allcock, S.; Young, E.H.; Holmes, M.; Gurdasani, D.; Dougan, G.; Sandhu, M.S.; Solomon, L.; Torok, M.E. Antimicrobial resistance in human populations: Challenges and opportunities. Glob. Health Epidemiol. Genom. 2017, 2, e4. [CrossRef] 8. Manresa, C.; Sanz-Miralles, E.C.; Twigg, J.; Bravo, M. Supportive periodontal therapy (SPT) for maintaining the dentition in adults treated for periodontitis. Cochrane Database Syst. Rev. 2018, 1, CD009376. [CrossRef] 9. Puri, K.; Puri, N. Local drug delivery agents as adjuncts to endodontic and periodontal therapy. J. Med. Life 2013, 6, 414–419. 10. H, R.R.; Dhamecha, D.; Jagwani, S.; Rao, M.; Jadhav, K.; Shaikh, S.; Puzhankara, L.; Jalalpure, S. Local drug delivery systems in the management of periodontitis: A scientific review. J. Control. Release 2019, 307, 393–409. [CrossRef] 11. Goodson, J.M. Gingival crevice fluid flow. Periodontol 2000 2003, 31, 43–54. [CrossRef][PubMed] Pharmaceutics 2020, 12, 352 12 of 14

12. Matesanz-Perez, P.; Garcia-Gargallo, M.; Figuero, E.; Bascones-Martinez, A.; Sanz, M.; Herrera, D. A systematic review on the effects of local antimicrobials as adjuncts to subgingival debridement, compared with subgingival debridement alone, in the treatment of chronic periodontitis. J. Clin. Periodontol. 2013, 40, 227–241. [CrossRef][PubMed] 13. Rasmussen, B.; Noller, H.F.; Daubresse, G.; Oliva, B.; Misulovin, Z.; Rothstein, D.M.; Ellestad, G.A.; Gluzman, Y.; Tally, F.P.; Chopra, I. Molecular basis of tetracycline action: Identification of analogs whose primary target is not the bacterial ribosome. Antimicrob. Agents Chemother. 1991, 35, 2306–2311. [CrossRef] [PubMed] 14. Castro, M.M.; Kandasamy, A.D.; Youssef, N.; Schulz, R. Matrix metalloproteinase inhibitor properties of tetracyclines: Therapeutic potential in cardiovascular diseases. Pharmacol. Res. 2011, 64, 551–560. [CrossRef] [PubMed] 15. Tomasi, C.; Wennstrom, J.L. Locally delivered doxycycline as an adjunct to mechanical debridement at retreatment of periodontal pockets: Outcome at furcation sites. J. Periodontol. 2011, 82, 210–218. [CrossRef] [PubMed] 16. Tonetti, M.S.; Lang, N.P.; Cortellini, P.; Suvan, J.E.; Eickholz, P.; Fourmousis, I.; Topoll, H.; Vangsted, T.; Wallkamm, B. Effects of a single topical doxycycline administration adjunctive to mechanical debridement in patients with persistent/recurrent periodontitis but acceptable oral hygiene during supportive periodontal therapy. J. Clin. Periodontol. 2012, 39, 475–482. [CrossRef] 17. Bassetti, M.; Schar, D.; Wicki, B.; Eick, S.; Ramseier, C.A.; Arweiler, N.B.; Sculean, A.; Salvi, G.E. Anti-infective therapy of peri-implantitis with adjunctive local drug delivery or photodynamic therapy: 12-month outcomes of a randomized controlled clinical trial. Clin. Oral Implants Res. 2014, 25, 279–287. [CrossRef] 18. Williams, R.C.; Paquette, D.W.; Offenbacher, S.; Adams, D.F.; Armitage, G.C.; Bray, K.; Caton, J.; Cochran, D.L.; Drisko, C.H.; Fiorellini, J.P.; et al. Treatment of periodontitis by local administration of minocycline microspheres: A controlled trial. J. Periodontol. 2001, 72, 1535–1544. [CrossRef] 19. OraPharma. ARESTIN Package Insert. 2017. Available online: https://www.orapharma.com/arestin (accessed on 13 April 2020). 20. Goodson, J.M.; Gunsolley,J.C.; Grossi, S.G.; Bland, P.S.; Otomo-Corgel, J.; Doherty,F.; Comiskey,J. Minocycline HCl microspheres reduce red-complex bacteria in periodontal disease therapy. J. Periodontol. 2007, 78, 1568–1579. [CrossRef] 21. Salvi, G.E.; Persson, G.R.; Heitz-Mayfield, L.J.; Frei, M.; Lang, N.P. Adjunctive local antibiotic therapy in the treatment of peri-implantitis II: Clinical and radiographic outcomes. Clin. Oral Implants Res. 2007, 18, 281–285. [CrossRef] 22. Kirchberg, M.; Eick, S.; Buchholz, M.; Kiesow, A.; Sarembe, S.; Mäder, K. Extrudates of lipophilic tetracycline complexes: A new option for periodontitis therapy. Int. J. Pharm. 2019, 572, 118794. [CrossRef][PubMed] 23. Mader, K.; Gallez, B.; Liu, K.J.; Swartz, H.M. Non-invasive in vivo characterization of release processes in biodegradable polymers by low-frequency electron paramagnetic resonance spectroscopy. 1996, 17, 457–461. [CrossRef] 24. Zhang, Y.; Schwendeman, S.P. Minimizing acylation of peptides in PLGA microspheres. J. Control. Release 2012, 162, 119–126. [CrossRef][PubMed] 25. Pirracchio, L.; Joos, A.; Luder, N.; Sculean, A.; Eick, S. Activity of taurolidine gels on ex vivo periodontal biofilm. Clin. Oral Investig. 2018, 22, 2031–2037. [CrossRef] 26. Kwasny, S.M.; Opperman, T.J. Static biofilm cultures of Gram-positive pathogens grown in a microtiter format used for anti-biofilm drug discovery. Curr. Protoc. Pharmacol. 2010.[CrossRef] 27. Pettit, R.K.; Weber, C.A.; Kean, M.J.; Hoffmann, H.; Pettit, G.R.; Tan, R.; Franks, K.S.; Horton, M.L. Microplate Alamar blue assay for Staphylococcus epidermidis biofilm susceptibility testing. Antimicrob. Agents Chemother. 2005, 49, 2612–2617. [CrossRef] 28. Budni, J.; Garcez, M.L.; de Medeiros, J.; Cassaro, E.; Bellettini-Santos, T.; Mina, F.; Quevedo, J. The anti-Inflammatory role of minocycline in Alzheimer s disease. Curr. Alzheimer Res. 2016, 13, 1319–1329. [CrossRef] 29. Brenes-Salazar, J.A. Minocycline: A bacteriostatic antibiotic with pleiotropic cardioprotective effects. Can. J. Physiol. Pharmacol. 2015, 93, 863–866. [CrossRef] Pharmaceutics 2020, 12, 352 13 of 14

30. Wang, X.; Ma, J.; Zhu, X.; Wang, F.; Zhou, L. Minocycline-loaded in situ hydrogel for periodontitis treatment. Curr. Drug Deliv. 2018, 15, 664–671. [CrossRef] 31. Jiang, X.; Liang, X.; Wang, S.; Guo, J.; Tao, Y.; Gui, S. An injectable in situ hexagonal mesophase system for local delivery of minocycline hydrochloride: Preparation and pharmacodynamics in rats. Pharmazie 2017, 72, 249–256. [CrossRef] 32. Yang, Z.; Liang, X.; Jiang, X.; Guo, J.; Tao, Y.; Wang, S.; Cao, Y.; Gui, S. Development and evaluation of minocycline hydrochloride-loaded in situ cubic liquid crystal for intra-periodontal pocket administration. Molecules 2018, 23, 2275. [CrossRef] 33. Roy, R.; Tiwari, M.; Donelli, G.; Tiwari, V. Strategies for combating bacterial biofilms: A focus on anti-biofilm agents and their mechanisms of action. Virulence 2018, 9, 522–554. [CrossRef][PubMed] 34. John, M.T.; Michalowicz, B.S.; Kotsakis, G.A.; Chu, H. Network meta-analysis of studies included in the Clinical Practice Guideline on the nonsurgical treatment of chronic periodontitis. J. Clin. Periodontol. 2017, 44, 603–611. [CrossRef][PubMed] 35. Scholz, M.; Reske, T.; Bohmer, F.; Hornung, A.; Grabow, N.; Lang, H. In vitro chlorhexidine release from alginate based microbeads for periodontal therapy. PLoS ONE 2017, 12, e0185562. [CrossRef][PubMed] 36. Zhang, Z.; Wang, Z.; Nong, J.; Nix, C.A.; Ji, H.F.; Zhong, Y. Metal ion-assisted self-assembly of complexes for controlled and sustained release of minocycline for biomedical applications. Biofabrication 2015, 7, 015006. [CrossRef][PubMed] 37. Kassem, A.A.; Ismail, F.A.; Naggar, V.F.; Aboulmagd, E. Comparative study to investigate the effect of meloxicam or minocycline HCl in situ gel system on local treatment of periodontal pockets. AAPS PharmSciTech 2014, 15, 1021–1028. [CrossRef] 38. Tew, J.G.; Marshall, D.R.; Burmeister, J.A.; Ranney, R.R. Relationship between gingival crevicular fluid and serum antibody titers in young adults with generalized and localized periodontitis. Infect. Immun. 1985, 49, 487–493. [CrossRef] 39. Brayton, J.J.; Yang, Q.; Nakkula, R.J.; Walters, J.D. An in vitro model of ciprofloxacin and minocycline transport by oral epithelial cells. J. Periodontol. 2002, 73, 1267–1272. [CrossRef] 40. Walters, J.D.; Nakkula, R.J.; Maney, P. Modulation of gingival fibroblast minocycline accumulation by biological mediators. J. Dent. Res. 2005, 84, 320–323. [CrossRef] 41. Sakellari, D.; Goodson, J.M.; Kolokotronis, A.; Konstantinidis, A. Concentration of 3 tetracyclines in plasma, gingival crevice fluid and saliva. J. Clin. Periodontol. 2000, 27, 53–60. [CrossRef] 42. Tonetti, M.; Cugini, M.A.; Goodson, J.M. Zero-order delivery with periodontal placement of tetracycline- loaded ethylene vinyl acetate fibers. J. Periodontal. Res. 1990, 25, 243–249. [CrossRef][PubMed] 43. Kim, T.S.; Klimpel, H.; Fiehn, W.; Eickholz, P. Comparison of the pharmacokinetic profiles of two locally administered doxycycline gels in crevicular fluid and saliva. J. Clin. Periodontol. 2004, 31, 286–292. [CrossRef] [PubMed] 44. Gentile, P.; Frongia, M.E.; Cardellach, M.; Miller, C.A.; Stafford, G.P.; Leggett, G.J.; Hatton, P.V. Functionalised nanoscale coatings using layer-by-layer assembly for imparting antibacterial properties to polylactide-co-glycolide surfaces. Acta Biomater. 2015, 21, 35–43. [CrossRef] 45. Kashi, T.S.; Eskandarion, S.; Esfandyari-Manesh, M.; Marashi, S.M.; Samadi, N.; Fatemi, S.M.; Atyabi, F.; Eshraghi, S.; Dinarvand, R. Improved drug loading and antibacterial activity of minocycline-loaded PLGA nanoparticles prepared by solid/oil/water ion pairing method. Int. J. Nanomedicine 2012, 7, 221–234. [CrossRef] [PubMed] 46. Klodzinska, S.N.; Wan, F.; Jumaa, H.; Sternberg, C.; Rades, T.; Nielsen, H.M. Utilizing nanoparticles for improving anti-biofilm effects of azithromycin: A head-to-head comparison of modified hyaluronic acid nanogels and coated poly(lactic-co-glycolic acid) nanoparticles. J. Colloid Interface Sci. 2019, 555, 595–606. [CrossRef][PubMed] 47. Han, C.; Goodwine, J.; Romero, N.; Steck, K.S.; Sauer, K.; Doiron, A. Enzyme-encapsulating polymeric nanoparticles: A potential adjunctive therapy in Pseudomonas aeruginosa biofilm-associated infection treatment. Colloids Surf. B Biointerfaces 2019, 184, 110512. [CrossRef][PubMed] 48. Gao, Z.; Song, M.; Liu, R.L.; Shen, Y.; Ward, L.; Cole, I.; Chen, X.B.; Liu, X. Improving in vitro and in vivo antibacterial functionality of Mg alloys through micro-alloying with Sr and Ga. Mater. Sci. Eng. C Mater. Biol. Appl. 2019, 104, 109926. [CrossRef] Pharmaceutics 2020, 12, 352 14 of 14

49. Nguyen, N.T.; Grelling, N.; Wetteland, C.L.; Rosario, R.; Liu, H. Antimicrobial activities and mechanisms of magnesium oxide nanoparticles (nMgO) against pathogenic bacteria, yeasts, and biofilms. Sci. Rep. 2018, 8, 16260. [CrossRef]

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).